AI Consulting & Engineering Automation · Zurich, Switzerland

AI-driven automation for teams shipping hardware.

Workflow automation and AI-assisted tooling, backed by deep embedded hardware, firmware, and PCB expertise for IoT teams, hardware startups, and product engineers — delivered remotely.

Discuss an AI Project View Services

Independent AI & engineering automation practice

AI ConsultingEngineering AutomationAI WorkshopsEmbedded Systems

When to engage

Engineering friction that slows teams down

Perri Engineering steps in when manual workflows or embedded complexity block product development. These are the problems that cost the most to leave unsolved.

Engineering workflows slowing iteration

Manual test processes, missing CI for firmware, or qualification pipelines that take days instead of hours. Automation and AI-assisted tooling can close these loops faster.

Unclear where AI actually fits

Interest in AI tooling without a clear picture of which parts of an engineering workflow it can safely improve — and which parts it should stay out of.

Embedded systems misbehaving

Prototypes that will not come up, intermittent firmware faults, or PCB issues surfaced after fabrication — problems where debug time gets expensive fast.

No embedded engineer in-house

Architecture decisions stalled, or a subsystem that needs specialist review before committing to fabrication, tooling, or production.

What we deliver

Engineering Services

AI-driven automation as the primary practice, backed by general embedded engineering consulting.

AI Consulting & Engineering Automation

Workflow automation and pragmatic AI tool adoption for engineering teams — automated verification pipelines, scripted tooling, and structured processes that remove manual bottlenecks from development.

When: Manual test cycles, slow verification loops, or a team evaluating where AI tooling can safely fit into an engineering workflow — and where it cannot.
Value: Faster iteration loops built on real engineering discipline, not a bolt-on tool — automated verification, scripted workflows, and grounded written recommendations.

AI Automation Workshops

A live, half-day session for your team — not a slide deck. Real examples of similar companies automating recognizable work, then a rough working prototype built live on a problem your team names.

When: A team curious about AI but unsure where it actually fits their workflow, and wants to see something built rather than pitched.
Value: A working prototype built in front of you, plus a written mini-audit specific to your case — the fee is credited in full toward a follow-on engagement booked within 30 days.

Embedded Engineering Consulting

General embedded consulting across hardware, firmware, and PCB work — board bring-up, bare-metal C/C++ and RTOS development, schematic and layout review, and structured root-cause analysis when systems misbehave.

When: Hardware will not come up, firmware faults resist conventional debugging, a layout needs review before fabrication, or an architecture decision needs an outside opinion.
Value: Documented root-cause findings and clear written recommendations, backed by lab measurement, schematic correlation, and systematic RCA methodology.

In practice

What this looks like day to day

Four everyday situations where automation and AI tooling replace manual engineering overhead.

The Friday report nobody wants to write

Today: Someone exports measurement logs, pastes them into a spreadsheet, and rebuilds the same status report by hand every week.
Automated: Collection and formatting run as a scripted step, so the report is an artifact the pipeline produces rather than an afternoon of someone’s week.

Test runs that only happen when someone remembers

Today: A change lands and the full suite is run manually, days later, if at all. Breakages surface long after the work that caused them.
Automated: The suite runs automatically on every change, so a regression is tied to the commit that introduced it instead of turning up weeks later.

Issue reports that arrive without enough detail

Today: Field reports and internal tickets come in unstructured. An engineer spends the first hour working out what was actually observed.
Automated: A triage step summarizes each report, groups related ones, and attaches the relevant logs before an engineer picks it up.

Knowledge that lives in one person’s inbox

Today: Decisions, measurements, and the reasoning behind them sit in chat threads and personal folders. The next engineer re-derives what was already solved.
Automated: Structured, searchable documentation is written as work happens, so context survives handovers and staff changes.

How I work

Your workflow first, then a solution

I start with how your team actually works today, then propose and build something tailored to that — not a generic template.

Study your workflow
A short call or visit to see how your team actually works today — the tools, handoffs, and manual steps — not a generic intake form.
Propose a tailored solution
A concrete recommendation for what to automate and how, specific to what was found — not a standard package sold to everyone.
Build and test
The automation, integration, or hardware review — delivered and verified. Working results, not just a proposal.
Handoff and support
Written findings, reproducible results, and documentation your team can build on. Follow-up support as the next steps progress.

Start the conversation

Ready to unblock your engineering team?

Describe the AI, automation, or hardware challenge you’re facing. Most engagements start with a brief diagnostic discussion — no commitment required.

Thomas S. PerriFounder, Perri Engineering